Drinking water is of importance for cancer studies attempting to measure both environmental and dietary exposures. Potentially useful data exist in water quality databases that have been created as part of routine, federally mandated surveillance activities. In many states these databases have existed for several decades, containing information about levels of contaminants such as nitrates, arsenic, or pesticides in drinking water supplies. Given the suspected long latent period between exposure and tumor diagnosis, or the potential importance of cumulative effects, these data may provide an opportunity to obtain measurements of exposures relevant to time periods prior to diagnosis for subjects in research studies. Our general purpose is to evaluate whether water contaminant levels obtained from an historic water surveillance database in Washington State can be used to estimate past and current individual exposure. We propose to focus on selected contaminants (nitrates and arsenic) that have been potentially associated with cancer occurrence. In Phase I of this project, geographic information systems methods (GIS) will be used to measure the correlation of tap water nitrate levels measured at residences of subjects enrolled in a previous cancer study, with nitrate levels measured in public water supplies from the same geographic coordinate. This will allow us to evaluate the extent to which it may be possible to measure prior exposure to drinking water contaminants using existing water quality surveillance data. In Phase 2, we will identify a new sample of residences located in regions where newly diagnosed cancer cases reside and measure tap water nitrate and arsenic levels, and conduct a similar evaluation to learn whether recent exposures may be estimated using the surveillance database. Phase 2 will also include an interview in which the level of tap water use (vs. bottled water or other source) for drinking, food preparation, and other modes of exposure are determined. If these data are correlated with tap water levels of contaminants, this method may be used to estimate previous and current exposures using methods that are less expensive and easier to employ in the context of epidemiologic studies. These methods may also be applicable to environmental databases with other types of exposures, or for examining other exposures (like pesticides) within drinking water surveillance databases.